Autonomous driving: Seeing is believing

Should we treat reports that ‘consumers don’t trust autonomous cars’, with suspicion?

When Google put 100 of its driverless test cars into the hands of its employees to use in their regular, everyday commute life, the company was surprised. Not by the fact that the cars did what they were supposed to, and none crashed. But that within a few hours of even the most hardened skeptic “Porsche driving enthusiast” had become so confident in the car’s autonomous capabilities that they started to exhibit "silly behaviour", performing other tasks and believing the car would cope with any situation and they could do over things. This despite being warned they needed to oversee the car and be prepared to take control at any given moment.

It was a story that resonated as I drove up the M1 Motorway last week in a new Mercedes E-Class equipped with the ‘Driving Assistance Plus’ package (the most important component of which is ‘Drive Pilot’). It means that the E-Class can, for surprisingly long periods of time, effectively drive itself.

It works so well on the motorway that after two hours I had completely bought in (the system tandems steer pilot, with the radar-guided cruise control and cameras which read the white lines and the speed limit signs to set its speed and keep distances). Taking your hands off the wheel of a fast-moving car feels un-natural at first and is something we are programmed through our driving life not to do, yet it takes a surprisingly short amount of time to trust the car.

Retrospectively, I found this odd. Despite being CDR’s resident leader on car-related technologies, and very 'pro-future car' I have been a sceptic of the autonomous hyperbole. Driverless cars on the very near horizon? The tech might be here, but we’re fighting against 100 years of learned behaviour (among car companies and consumers), our crumbling infrastructure, the sloth-like rate of progress and risk aversion of legislators and insurers, and the mature driver attitude, which — if we consider the average Western consumer — doesn’t believe that a computer and software can actually do the job of driving a car better than they can.

Plus, as everyone who works at CDR will tell you, I break technology. New iPhone 6? Dead in my hands after three weeks. 2016 MacBook Pro? Replaced under warranty by Apple at two months old. If it’s software dependent, I have a worrying habit of killing it.

So letting 50-grands worth of very new Mercedes pilot itself down a UK motorway is — considered here sat at my desk — not something I’d be happy with.  What I’d tell you if you interviewed me on the subject, is that I’d be driving it myself. Switch off all the assist systems, just concentrate on the road and drive. 

The reality? I leave Silverstone race track in the Merc on Thursday afternoon. Spend the first half hour playing around with the impressive — but (on first acquaintance) fiercely complex —interface, as I drive (horribly distracted). Get to the motorway, 5pm, predictably busy. Join the motorway, move into the fast lane. Traffic flow speed is oscillating between about 50-75 mph. Switch the cruise control on. A display appears in the digital cluster, showing a car in front — speed is set to 70 mph. Feet off the pedals, car now maintains a distance to the car in front. So far, so normal. A mid-range Golf has this tech. Then feel the steering tugging slightly, occasionally against my inputs. Note the little green steering wheel symbol on the IP is illuminated. Let go of the wheel on a straight bit, and realise that the car is constantly positioning itself in the centre of the lane. Decide, as a game, to see how long I can keep my hands off the wheel without it freaking out (answer, about 20 seconds — until a symbol appears instructing me to hold the wheel). Find the car steers round motorway-radius curves quite comfortably. Resume game, arbitrarily touching the steering wheel about every 20 seconds. Decide to extend the game, see how long I can go without needing to do anything except occasionally touching the wheel to let the car know I’m still there.

An hour later, realise I’ve still not touched the pedals. Car slows to the back of a traffic jam. Traffic stops. The car in front moves off, E-Class follows. As the traffic jam creeps along, I realise it’s stopped asking me to touch the wheel. Carry on in this vain for about 5 miles. Phone my friend to tell him about how amazing this is. Finish call. Car still happily doing this itself. Start doing silly things like holding my hands up and reclining the seat, to try and test the reaction of drivers around me and see if they notice effectively no one is driving. No one does. Two hours in, I realise that — far from being upset at any notion of lost control on my part, I now completely trust the car, and am seriously considering getting my phone out of the centre bin, to start replying to emails rather than wasting time in this traffic jam. 

It’s a story to illustrate how, when we see headlines like the recent ‘half of motorists oppose driverless cars’, we might consider that research somewhat unreliable. Most people today haven’t experienced Level 2 autonomy (how the Merc is classed) — let alone full — autonomous driving capabilities yet. But when they do, we might be surprised at just how quickly they become completely comfortable with it, like I did. Ultimately, seeing is believing. 

Posted by Joe Simpson


Tags

Quarterly Insights

Sign up to receive our quarterly review of what's happening in automotive design


Archive

  1. 2017
    1. June
    2. May
    3. April
    4. March
  2. 2016
    1. November
    2. September
    3. July
    4. June
    5. May
    6. April
    7. March
    8. February
    9. January
  3. 2015
    1. December
    2. November
    3. October
    4. September
    5. August
    6. July
    7. June
    8. May
    9. April
    10. March
    11. February
    12. January
  4. 2014
    1. December
    2. November
    3. October
    4. September
    5. August
    6. July
    7. May
    8. April
    9. March
    10. February
    11. January
  5. 2013
    1. December
    2. September
    3. June

We help our clients to design more successful cars.

info@cardesignresearch.com